Ukraine’s Zelenskiy says Russia has readied 30,000 North Korean
Zelenskiy says Russia is preparing 30,000 North Korean troops for Voronezh deployment since June, raising geopolitical risks with crypto market
TL;DR
North Korean group Kimsuky has begun using local AI tools to automate cyber attacks, while the U.S. DOJ charged four North Koreans with stealing nearly $1 million in crypto from blockchain firms. The moves show a shift toward AI‑driven threats and heightened enforcement against illicit crypto funding.
Key facts
- Kimsuky installed AI models Ollama, GPT‑4All and Msty on its own infrastructure.
- Four North Korean nationals were charged with stealing $915,000 in crypto.
- AI‑enabled scam activity rose about 500 percent in the last year.
- Hackers caused $972 million in losses in H1 2026, down from $2.3 billion in H1 2025.
North Korean cyber‑espionage group Kimsuky has begun installing local artificial‑intelligence tools to automate its attacks, while the U.S. Department of Justice has charged four North Korean nationals with stealing nearly $1 million in cryptocurrency from blockchain firms. The AI deployment was reported by South‑Korea‑based firm Genians on July 26, 2026, and the DOJ indictment, also released on July 26, names Kim Kwang Jin, Kang Tae Bok, Jong Pong Ju and Chang Nam Il as the accused. Both developments highlight a shift toward more sophisticated, AI‑driven operations and continued illicit fundraising for the Pyongyang regime.
Background
Kimsuky, a cyber‑espionage outfit linked to North Korea, traditionally targets governments, researchers and businesses. According to Genians, the group has begun installing AI models such as Ollama, GPT‑4All and Msty on its own infrastructure, allowing it to analyse stolen documents without sending data to external services. The firm also observed the use of retrieval‑augmented generation (RAG) technology to quickly search large data sets, speech‑to‑text software, the Cursor coding assistant and generative AI for crafting convincing phishing content. In parallel, the U.S. Department of Justice’s DPRK RevGen: Domestic Enabler Initiative, launched in 2024, seeks to disrupt North Korea’s illicit revenue streams. The DOJ’s recent indictment alleges that four North Korean operatives posed as remote developers for U.S. and Serbian blockchain companies, using fabricated identities to gain privileged access and siphon crypto assets, which were then laundered through mixers and exchange accounts. These two strands illustrate a broader pattern: North Korean actors are expanding their technical toolkit while continuing to exploit the global crypto ecosystem for funding.
What happened
On July 26, 2026, Genians disclosed that Kimsuky had installed multiple AI tools on its own servers, including Ollama, GPT‑4All and Msty, to examine stolen documents without exposing data to public AI services. The firm said the group also deployed RAG technology to search and extract information from large stolen file collections, used speech‑to‑text software, and employed the Cursor AI coding tool to accelerate malware development. Generative AI was further used to produce realistic phishing documents, making scams harder to detect. In the same day’s reporting, the U.S. Department of Justice announced charges against four North Korean nationals—Kim Kwang Jin, Kang Tae Bok, Jong Pong Ju and Chang Nam Il—for wire fraud and money laundering after they posed as remote IT workers at an Atlanta‑based blockchain startup and a Serbian token firm. Prosecutors detailed that the defendants used stolen and fabricated IDs to secure positions, then stole about $915,000 in crypto: roughly $175,000 in February 2022 and $740,000 in March 2022, moving the proceeds through mixers and exchange accounts set up with fraudulent Malaysian IDs. The DOJ linked the scheme to the DPRK RevGen: Domestic Enabler Initiative, which targets North Korean revenue‑generating activities. Both sources underscore a coordinated effort by North Korean actors to leverage advanced technology and deceptive employment practices to fund illicit programs.
Market impact
TRM Labs reported that AI‑enabled scam activity rose about 500 percent over the previous year, while DeFiLlama data showed total value hacked in 2026 reaching $1.195 billion, down from $2.3 billion in the first half of 2025. In the first half of 2026, hackers executed a record 207 hacks causing $972 million in losses, a figure less than half of the $2.3 billion taken in the same period of 2025. The DOJ case involved the theft of nearly $1 million in crypto, which was laundered through mixers and exchange accounts before being seized. These numbers illustrate a contraction in total crypto theft volume but a sharp increase in AI‑driven scams, suggesting a shift in attacker tactics that could affect investor confidence and the demand for security solutions in the DeFi space.
Why it matters
The integration of AI tools by Kimsuky indicates that North Korean cyber actors are moving toward more automated, efficient attacks that reduce reliance on external services, potentially lowering detection risk. The DOJ’s indictment demonstrates that the United States is intensifying legal pressure on North Korean cyber‑criminals, targeting both the operatives and the financial pathways they use. Together, these developments suggest a dual pressure on the crypto ecosystem: heightened technical sophistication of threats and increased enforcement actions. For blockchain projects and investors, the rise of AI‑enhanced phishing and malware could raise operational costs for security, while the legal crackdown may deter some illicit actors but also push them to adopt even more covert methods.
DeFi Intel’s take
Editorial analysis, not a statement of fact.
DeFi Intel's take: the convergence of AI adoption within Kimsuky’s toolkit and the DOJ’s aggressive prosecution of North Korean crypto thieves suggests the threat landscape is evolving on two fronts. We read this as a signal that state‑linked actors are seeking to outpace traditional defenses by automating reconnaissance and phishing, while law‑enforcement is attempting to cut off the financial lifelines that fund these operations. The 500 percent surge in AI‑enabled scams, paired with a drop in overall hack value, hints that attackers are focusing on higher‑success, lower‑volume tactics. Projects that invest early in AI‑aware security measures may gain a competitive edge as the threat vector matures.
Sources used for this analysis
- ambcrypto.com/north-korean-hackers-turn-to-ai-as-crypto-scams-surge-report/
- cointelegraph.com/news/north-korean-it-workers-charged-crypto-theft
DeFi Intel wrote this analysis from the reporting above. Any figure or quotation can be checked against those pages.
Frequently asked questions
What AI tools has the Kimsuky group installed?
According to Genians, Kimsuky installed local AI models such as Ollama, GPT‑4All and Msty, along with retrieval‑augmented generation technology, speech‑to‑text software, the Cursor coding assistant and generative AI for phishing content.
How much cryptocurrency did the four North Korean defendants steal?
The Department of Justice alleges the four defendants stole about $915,000 in crypto, with roughly $175,000 taken in February 2022 and $740,000 in March 2022, before laundering the proceeds through mixers and exchange accounts.
What trend in crypto scams does TRM Labs report?
TRM Labs reports that AI‑enabled scam activity increased by about 500 percent over the past year, indicating a rapid rise in the use of artificial intelligence by cyber‑criminals.
Direct answers
Short answers, written to be quoted.
- What AI tools are North Korean hackers using?
- North Korean hackers are using local AI tools such as Ollama, GPT‑4All, Msty, retrieval‑augmented generation technology, speech‑to‑text software and the Cursor coding assistant.
- How much crypto did the DOJ say was stolen by the North Korean defendants?
- The DOJ says the four defendants stole about $915,000 in cryptocurrency, with $175,000 taken in February 2022 and $740,000 in March 2022.
DeFi Intel is an entity-graph aggregator: we curate, tag and link crypto news to a typed knowledge graph of protocols, tokens, people and incidents. We do not republish the full article body. Use the link above to read the original report at Crypto Briefing.
Discussion
Comments are reviewed before publication.
Want the full article?
Continue reading on Crypto Briefing →
No comments yet.